(MIT Introduction to Computational Thinking and Data Science Series)

Note: how to judge whether a sample is random walk or not.
Note: Law of large numbers – Bernoulli Law

Gambler’s fallacy: misunderstanding of LOLN. Deviations from mean will even out in the future.
Story: Roulette gave 25 times of black consecutively. People came to bet on red as they thought things got to even out. — below the mean.


Parents who are bigger than average will have children who are smaller than average. Follow an extreme event, the next event will be less extreme.
The more sample you take, the more likely you will get to the mean.


confidence interval: empirical rule with assumptions


Note: As confidence interval shrinks, we are sure that our mean is true mean.
Assumption of ER: Mean estimation error is 0. We are same likely to guess high and to guess low. Labortoary might have bias.



